{"id":"https://openalex.org/W7204128790","doi":"https://doi.org/10.48550/arxiv.2608.21024","title":"RODE: A Radial-Orthogonal Decoupled Engine for Optimization","display_name":"RODE: A Radial-Orthogonal Decoupled Engine for Optimization","publication_year":2026,"publication_date":"2026-08-21","ids":{"openalex":"https://openalex.org/W7204128790","doi":"https://doi.org/10.48550/arxiv.2608.21024"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.21024","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21024","pdf_url":"https://arxiv.org/pdf/2608.21024","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.21024","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5147935358","display_name":"Guoxiang Xu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xu, Guoxiang","raw_affiliation_strings":["Zhejiang University , Hangzhou , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University , Hangzhou , China","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148023361","display_name":"Bince Qu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Qu, Bince","raw_affiliation_strings":["Zhejiang University , Hangzhou , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University , Hangzhou , China","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5147928303","display_name":"Qi Sun","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Sun, Qi","raw_affiliation_strings":["Zhejiang University , Hangzhou , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University , Hangzhou , China","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5148011780","display_name":"Cheng Zhuo","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhuo, Cheng","raw_affiliation_strings":["Zhejiang University , Hangzhou , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University , Hangzhou , China","institution_ids":["https://openalex.org/I76130692"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I76130692"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9897000193595886,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9897000193595886,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.0013000000035390258,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.0013000000035390258,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/muon-optimizer","display_name":"Muon optimizer","score":0.6259999871253967},{"id":"https://openalex.org/keywords/llm-fine-tuning","display_name":"LLM fine-tuning","score":0.7990000247955322},{"id":"https://openalex.org/keywords/image-classification","display_name":"image classification","score":0.7300000190734863},{"id":"https://openalex.org/keywords/gpt-2","display_name":"GPT-2","score":0.5320000052452087}],"concepts":[{"id":"https://openalex.org/C191795146","wikidata":"https://www.wikidata.org/wiki/Q3878446","display_name":"Norm (philosophy)","level":2,"score":0.6582000255584717},{"id":"https://openalex.org/C205606062","wikidata":"https://www.wikidata.org/wiki/Q5249645","display_name":"Decoupling (probability)","level":2,"score":0.5310999751091003},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.499099999666214},{"id":"https://openalex.org/C47446073","wikidata":"https://www.wikidata.org/wiki/Q5165890","display_name":"Control theory (sociology)","level":3,"score":0.49779999256134033},{"id":"https://openalex.org/C138187205","wikidata":"https://www.wikidata.org/wiki/Q131251","display_name":"Tangent","level":2,"score":0.4163999855518341},{"id":"https://openalex.org/C57691317","wikidata":"https://www.wikidata.org/wiki/Q1289248","display_name":"Scalar (mathematics)","level":2,"score":0.3865000009536743},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.3564000129699707},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.34459999203681946},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.3424000144004822},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3165999948978424},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2833000123500824},{"id":"https://openalex.org/C92207270","wikidata":"https://www.wikidata.org/wiki/Q939253","display_name":"Matrix norm","level":3,"score":0.2809000015258789},{"id":"https://openalex.org/C146324458","wikidata":"https://www.wikidata.org/wiki/Q1202673","display_name":"Uniform norm","level":2,"score":0.26190000772476196},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2565999925136566}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.21024","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21024","pdf_url":"https://arxiv.org/pdf/2608.21024","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.21024","doi":"https://doi.org/10.48550/arxiv.2608.21024","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.21024","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2608.21024","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21024","pdf_url":"https://arxiv.org/pdf/2608.21024","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204128790.pdf","grobid_xml":"https://content.openalex.org/works/W7204128790.grobid-xml"},"referenced_works_count":15,"referenced_works":["https://openalex.org/W1483804921","https://openalex.org/W2108598243","https://openalex.org/W2138674039","https://openalex.org/W2626017178","https://openalex.org/W2757910899","https://openalex.org/W2963248893","https://openalex.org/W3210904070","https://openalex.org/W4211115884","https://openalex.org/W4401660192","https://openalex.org/W4414846785","https://openalex.org/W4415796274","https://openalex.org/W4415798387","https://openalex.org/W7101563876","https://openalex.org/W7164920391","https://openalex.org/W7165948081"],"related_works":["https://openalex.org/W7165948081","https://openalex.org/W7165626314","https://openalex.org/W7160960074","https://openalex.org/W4415089271","https://openalex.org/W7128499965","https://openalex.org/W7134925487","https://openalex.org/W7166508516","https://openalex.org/W7160875907","https://openalex.org/W7160416436","https://openalex.org/W7140217466"],"abstract_inverted_index":{"Modern":[0],"neural":[1],"network":[2],"training":[3],"increasingly":[4],"uses":[5],"matrix-aware":[6],"optimizers,":[7],"yet":[8],"their":[9],"conditioned":[10],"matrix":[11,68,213],"step":[12,62],"is":[13],"typically":[14],"added":[15],"directly":[16,131],"to":[17,142,150,153,212],"the":[18,30,53,67,84,127,133,169,174,192],"weight,":[19],"jointly":[20],"changing":[21],"its":[22,77],"norm":[23,32,41,70,95,147],"and":[24,43,55,61,97,104,117,144,173,177,190,197,202,209],"direction.":[25],"This":[26],"interaction":[27],"matters":[28],"because":[29],"current":[31],"determines":[33],"angular":[34],"motion,":[35],"while":[36,76],"directional":[37,56,78,99,203],"learning":[38,128],"can":[39],"drive":[40],"growth":[42],"thereby":[44],"alter":[45],"later":[46],"steps.":[47],"We":[48],"introduce":[49],"RODE,":[50],"which":[51],"gives":[52],"radial":[54,74,201],"components":[57],"separate":[58],"update":[59],"rules":[60],"sizes.":[63],"RODE":[64,108,137,158,180],"explicitly":[65],"updates":[66,82],"Frobenius":[69],"through":[71],"a":[72,206],"scalar":[73],"rule,":[75],"channel":[79],"performs":[80],"Newton--Schulz-conditioned":[81],"in":[83,113],"tangent":[85],"space.":[86],"Controlled":[87],"GPT-2":[88],"interventions":[89],"show":[90],"gains":[91],"from":[92,132,140,148],"both":[93,110,182],"direct":[94,115],"control":[96],"RODE's":[98],"update.":[100],"Across":[101],"two":[102,105],"language-modeling":[103],"image-classification":[106],"tasks,":[107],"outperforms":[109,181],"Muon":[111,154,183],"variants":[112,184],"every":[114],"comparison":[116],"ends":[118],"with":[119,156],"lower":[120],"full-model":[121],"norms.":[122],"At":[123],"1.5B":[124],"scale,":[125],"using":[126],"rate":[129],"transferred":[130],"Qwen2-style":[134],"LM":[135],"sweep,":[136],"lowers":[138],"loss":[139],"4.145":[141],"3.346":[143],"final":[145],"global":[146],"11964":[149],"2183":[151],"relative":[152],"RMS,":[155],"fixed-radius":[157],"improving":[159],"further.":[160],"For":[161],"Qwen3.5-9B":[162],"full-parameter":[163],"fine-tuning,":[164],"all":[165,186],"six":[166],"optimizers":[167],"use":[168],"same":[170,175],"tuning":[171],"budget":[172],"formal-training":[176],"evaluation":[178,188],"settings;":[179],"on":[185,195],"four":[187],"tasks":[189],"attains":[191],"highest":[193],"mean":[194],"GSM8K":[196],"MATH-500.":[198],"Thus,":[199],"decoupling":[200],"dynamics":[204],"offers":[205],"more":[207],"effective":[208],"controllable":[210],"approach":[211],"optimization.":[214]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-08-25T00:00:00"}